文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > member of the International Association of Geodesy (IAG) Special Study Group (SSG) 4.190 No
侵权投诉

member of the International Association of Geodesy (IAG) Special Study Group (SSG) 4.190 No

(UNSW). His current research interests are in modeling, quality control, algorithm and software development for precise positioning using GPS, Glonass, pseudolites and INS. He holds a Ph.D. from the Curtin University of Technology, Australia, and has autho

level system, i=1,2,..mwhere m is the number ofantennas minus one, for each pair of antennas, thefollowing relationship is obtained:

ei=Rbebi

, i=1,2,..m (2)

If only two antennas are amounted on the platform, one

baseline vector is available to estimate two-dimensionalattitude, pitch and yaw or roll and yaw, depending on thearrangement of the antennas. If two or more baselinevectors are available, three-dimensional attitudeparameters can be determined, and there are also morethan three redundant measurements. In all scenarios, aleast-squares procedure can be used to estimate theparameters.

Linearizing equation (2) with respect to the attitudeparameters (φ,θ and ψ), the design matrix H isobtained, in which all the components are the partialderivatives calculated using the initial values for theattitude parameters. Furthermore, if the covariance matrixfor the GPS-derived baseline vectors (ei) is given as

σ Q, where σ is the standard deviation of the GPS

measurements and Q is the cofactor matrix, thecovariance matrix for the estimated attitude parameters is

C=σ (HTQ 1H)

1

(3)

The accuracy of the estimated attitude parameters can beobtained from the diagonal elements of the matrix C.Similar to the DOP values for GPS single-point pointing,the DOP values for the attitude parameters (φ,θ and ψ),i.e., DOP(φ), DOP(θ) and DOP(ψ), can be calculatedusing equation (3).

From the above discussion it is clear that reliability andaccuracy of the estimated attitude parameters is heavilydependent on the quality of the GPS-derived baselinevectors, which can be improved by adding the pseudolitemeasurements into the solution.THEORETICAL ANALYSIS

From the positioning point of view, a pseudolite can beconsidered to be a ‘satellite-on-the ground’ (Hein et al.,1997). The pseudolites can also transmit GPS-likesignals. Therefore, the fundamental principles in preciseGPS positioning, such as interferometry, can be applied tointegrated GPS-pseudolite positioning. However, the useof ground-based pseudolites does have differentimplications for some aspects of positioning operations.For example, for short baselines as used in attitudedetermination, the effects of nonlinearity and orbital errorson the measurement models are negligible for thesatellites, but could be significant in the case ofpseudolites. The reason for this is that pseudolites are

much closer to user receivers than the satellites are. A

closer investigation of these issues has been made.Effects of Nonlinearity

In positioning and navigation applications, the keygeometric information from the measurements is 'distance'or 'range' between two points, which are generallyrepresented through 'coordinates' defined in somereference frame. The relationship between themeasurements (distances) and the unknown parameters(coordinates) is of course nonlinear. Because theestimation techniques for the linear models have nicestatistical properties, the nonlinear measurement modelsare usually linearized using a Taylor series expansion.A function F(x) is expanded into a Taylor series up tothe first order as

F(x)=F(x0)+Aδx+R

(4)

where x0 is the approximate values for the parameters x;A= xF(x0) is the vector of first order partialderivatives evaluated at x0, which is also called the line-of-sight (LOS) vector; δx=x x0; and R is the secondorder remainder term expressed as

R=12

δxT 2

xxF(x0+t δx)δx, 0<t<1

(5)

where 2xxF(x) is the Hessian matrix constructed usingthe second order partial derivatives.

In standard data processing, the remainder term R isignored, resulting in a nonlinearity error in themeasurement model. The bounds of this error term aregiven as (Teunissen, 1987):

12λ2R≤12minx≤2

λmaxx (6)

where λmin and λmax are the minimum and maximumeigenvalues of the Hessian matrix 2xxF(x). For thedistance model:

dij=xij2+yij2+z2ij

, (7)

where xij,yij,zij are the differences between the

coordinates for the two points (i,j) associated with themeasurement, the Hessian matrix is

上一页第3页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email